chore: prepare config versions for release

Drop unused process metrics and disable root cgroup stats (host-wide
totals already come from prometheus.exporter.unix).

.env.sample

@@ -20,9 +20,6 @@ SECRET_BASIC_AUTH_VERSION=v1
 # server is remote
 # PROMETHEUS_REMOTE_WRITE_URL=https://prometheus.$DOMAIN/api/v1/write
 
-# Enable container filesystem usage metrics (expensive du-scan, ~30% CPU on large hosts)
-# CADVISOR_DISK_USAGE=1
-
 # Enable authenticated scraping of containers that opt in via
 # prometheus.io/auth=basic or prometheus.io/auth=bearer labels (used as
 # password/bearer token respectively). Insert it with:
@@ -87,7 +84,7 @@ SECRET_BASIC_AUTH_VERSION=v1
 # SECRET_GF_ADMINPASSWD_VERSION=v1
 ## Grafana's own domain. Defaults to $DOMAIN
 ## Change the value if you want Grafana on another domain.
-# GRAFANA_DOMAIN=$DOMAIN
+GRAFANA_DOMAIN=$DOMAIN
 #
 ## Single-Sign-On with OIDC
 # COMPOSE_FILE="$COMPOSE_FILE:compose.grafana-oidc.yml"

abra.sh

@@ -10,7 +10,7 @@ export PROMETHEUS_YML_VERSION=v2
 export MATRIX_ALERTMANAGER_CONFIG_VERSION=v1
 export MATRIX_ALERTMANAGER_ENTRYPOINT_VERSION=v1
 export GF_ALERTS_NODE_VERSION=v3
-export CONFIG_ALLOY_VERSION=v3
+export CONFIG_ALLOY_VERSION=v1
 
 # migrates secrets from old names to new names by reading values from the
 # running containers on the server and re-inserting them under the new names.

compose.yml

@@ -53,7 +53,7 @@ services:
         condition: on-failure
       labels:
         - "backupbot.backup=${ENABLE_BACKUPS:-true}"
-        - "coop-cloud.${STACK_NAME}.version=2.0.0+v1.18.1"
+        - "coop-cloud.${STACK_NAME}.version=1.6.0+v1.8.1"
 configs:
   config_alloy:
     template_driver: golang

config.alloy.tmpl

@@ -14,11 +14,7 @@ discovery.docker "linux" {
 {{ if ne (env "PROMETHEUS_REMOTE_WRITE_URL") "" }}
 prometheus.exporter.cadvisor "docker" {
   docker_only = true
-  {{ if eq (env "CADVISOR_DISK_USAGE") "1" }}
   enabled_metrics = ["cpu", "cpuLoad", "disk", "diskIO", "memory", "network"]
-  {{ else }}
-  enabled_metrics = ["cpu", "cpuLoad", "diskIO", "memory", "network"]
-  {{ end }}
   // host-wide totals already come from prometheus.exporter.unix
   disable_root_cgroup_stats = true
 }
@@ -52,32 +48,7 @@ prometheus.scrape "default" {
     prometheus.exporter.cadvisor.docker.targets,
   )
 
-  forward_to = [prometheus.relabel.container_meta.receiver]
-}
-
-prometheus.relabel "container_meta" {
   forward_to = [prometheus.remote_write.prometheus.receiver]
-
-  // remove sha tail: nginx:1.31.1@sha256:608a... -> nginx:1.31.1
-  rule {
-    source_labels = ["image"]
-    regex         = "([^@]+)@sha256:.*"
-    target_label  = "image"
-    replacement   = "$1"
-  }
-  // split image in name and tag
-  rule {
-    source_labels = ["image"]
-    regex         = "(.+):[^:/]+"
-    target_label  = "image_name"
-    replacement   = "$1"
-  }
-  rule {
-    source_labels = ["image"]
-    regex         = ".+:([^:/]+)"
-    target_label  = "image_tag"
-    replacement   = "$1"
-  }
 }
 
 prometheus.remote_write "prometheus" {
@@ -315,35 +286,7 @@ loki.source.docker "docker" {
 loki.source.journal "journal" {
   path = "/rootfs/var/log/journal"
   labels = { job = "{{ env "DOMAIN" }}" }
-  relabel_rules = loki.relabel.journal.rules
-  forward_to = [loki.process.journal.receiver]
-}
-
-loki.relabel "journal" {
-  forward_to = []
-
-  rule {
-    source_labels = ["__journal__systemd_unit"]
-    target_label  = "unit"
-  }
-  rule {
-    source_labels = ["__journal_syslog_identifier"]
-    target_label  = "ident"
-  }
-  rule {
-    source_labels = ["__journal_priority_keyword"]
-    target_label  = "priority"
-  }
-}
-
-// drop network db stats (15-20% of kernel logs)
-loki.process "journal" {
   forward_to = [loki.write.loki.receiver]
-
-  stage.drop {
-    expression          = ".*NetworkDB stats.*"
-    drop_counter_reason = "networkdb_stats"
-  }
 }
 {{ end }}
 

release/next

@@ -59,6 +59,3 @@ See the README's "Auto-discovering metrics from other apps" section.
 The Swarm, Stacks and Traefik dashboards were reworked to show old (pull-model)
 and new (Alloy push-model) data as one continuous line, so you don't lose history
 across the migration.
-
-The disk usage per container was disabled per default because of the CPU-expensive 
-filesystem scan (30% cpu increase). Enable by uncommenting env CADVISOR_DISK_USAGE=1